ABSTRACT Delay in admission to inpatient rehabilitation facilities (rehabilitation latency) after stroke is associated with poor health outcomes. The objectives of this proposal are to (a) identify factors associated with rehabilitation latency to inpatient rehabilitation facilities (IRF), and (b) compare health outcomes associated with low versus high rehabilitation latency post-acute stroke. This observational study will use secondary analyses of Medicare claims data. We will utilize the Centers for Medicare and Medicaid data files (Medicare Provider Analysis and Review, Beneficiary Summary, Inpatient Rehabilitation Facilities Patient Assessment Instrument, and Provider of Service) for the years 2016-2017. Patients with ?index stroke? admitted directly to IRFs from acute hospitals will serve as our main cohort. For the first aim, rehabilitation latency will serve as the primary outcome, adjusting for patient-, hospital-, and market-level characteristics. For the second aim, rehabilitation latency will be the primary independent variable for estimating health outcomes, namely: functional gain, successful community discharge after inpatient rehabilitation, home-to-home time, readmission to acute hospitals, and six-month mortality. We will match patients by comorbid conditions and hospital-related complications in order to reduce selection bias in our sample, using propensity score methods. Given the presence of nesting effects, we plan to use hierarchical generalized linear and linear mixed models for the outcomes. With this proposal, we anticipate exploring hospital-related factors as one set of mechanisms behind rehabilitation latency into IRFs. Furthermore, we expect to see significant association between rehabilitation latency and post-IRF outcomes (functional gain, and successful discharge to home), and no association with post-IRF discharge outcomes (rehospitalization and mortality).